The index is at 22,000, the one-month future at 22,100 and the two-month future at 22,300, with one month of carry worth 0.47%. Which leg is mispriced, what calendar spread follows, and what can go wrong before expiry?

Fair values are about 22,103 and 22,207, so the near month is about fair and the next month at 22,300 is 93 points rich. Sell the next month and buy the near month: the spread is quoted at 200 against a fair 104. If carry is unchanged when the near month expires, the spread collapses to about 103 and 100 lots of each make about Rs 4.8 lakh. The trade is exposed to carry, rates and dividend expectations, not to the index level.

A shop that sells you rice for delivery next month should charge today's price plus what it costs to hold the rice for a month: the shop's interest on the money, less any income the rice earns. A futures price on an index is the same calculation: spot plus the cost of financing the stocks, less the dividends they pay, which here is 0.47% a month, so fair values are 22,000 times 1.0047, 22,103.4, and 22,000 times 1.0047 squared, 22,207.3. Against those, the near month at 22,100 is 3 points cheap, which is noise, and the next month at 22,300 is 93 points rich, which is not. The calendar spreadA position long one futures expiry and short another on the same underlying, which gains or loses on the price gap between them rather than on the level of the underlying. between them is quoted at 200 points against a fair 104.

Selling the rich next-month future alone would be a bet that the index falls. Buying the near month against it cancels the index exposure, so the position makes money if the gap between the two months shrinks from 200 points towards its fair value, whatever the index does. When the near month expires it equals spot, and the next month, now one month from expiry, should be worth spot plus one month of carry. If the index is still 22,000 and carry still 0.47%, the gap is about 103 points, so the spread has fallen by 96.6 points; on 100 lots of 50 units each side that is about Rs 4.8 lakh. The table shows how the result depends on carry far more than on the index.

Ask first why the next month might be at 22,300 for a reason. The quotes imply 0.90% of carry for the second month, nearly double the desk's 0.47%, so either the market expects much higher financing costs, or much lower dividends in that month, or the desk's carry estimate is wrong; if any of those is true, the trade is not an arbitrage but a bet against better information. A large company skipping a dividend expected in month two, or a sharp rise in short-term rates, would lift the fair spread and erode the gain. There are also mechanical risks: both legs are margined, so the desk posts margin on two positions; the near month expires first, leaving the short next-month leg naked unless it is rolled or closed at the same time; and a spike in demand for the next month around an index event can widen the spread further before it narrows, which is a mark-to-market loss even if the trade works in the end.

Say the limits. The 0.47% is an estimate built from the desk's financing rate and expected dividends, and that estimate is the whole trade; transaction costs and the bid-offer on both legs are ignored; and the arbitrage that would truly lock the 93 points, selling the next month and buying the index basket, is avoided here because the desk does not want to hold stocks. The judgement: the spread is wide enough to trade at modest size, after checking the dividend calendar for month two.